143

Molecular Pharmacology of Agouti Protein

in Vitro

and

in Vivo

GREGORY S. BARSH,

a

MICHAEL M. OLLMANN, BRENT D. WILSON, KIMBERLY A. MILLER, AND TERESA M. GUNN

Departments of Pediatrics and Genetics, and the Howard Hughes Medical Institute, Stanford University School of Medicine, Stanford, California, USA

ABSTRACT: Agouti protein and Agouti-related protein (Agrp) are paracrinesignaling molecules that act by antagonizing the effects of melanocortins, andseveral alternatives have been proposed to explain their mechanisms of action.Genetic crosses in a sensitized background uncover a phenotypic differencebetween overexpression of Agouti and loss of Mc1r function, demonstrate thata functional Mc1r is required for the pigmentary effects of Agouti, and suggestthat Agouti protein can act as an agonist of the Mc1r in a way that differsfrom

�

-MSH stimulation.

In vitro,

Agouti protein inhibits melanocortin actionby two mechanisms: competitive antagonism that depends on the carboxy-terminus of the protein, and downregulation of melanocortin receptor signal-ing that depends on the aminoterminus. Our findings provide evidence of anovel signaling mechanism whereby

�

-MSH and Agouti protein function as in-dependent ligands that inhibit each other’s binding and transduce opposite sig-nals through a single receptor.

INTRODUCTION

The genetics of mouse coat color offers a model system for studying gene actionand interaction. There are nearly 100 known, or so-called classical, coat color muta-tions; many of which have been cloned and characterized.

1,2

The relative accessibil-ity of skin and hair to experimental manipulation allows one to determine the celltype in which a particular gene is required. Finally, the ability to generate and ana-lyze double mutants permits the construction of a genetic framework to build bio-chemical and cell-biologic pathways.

Studies carried out in our laboratory and those of others over the past decade havedemonstrated that the mouse

Agouti

gene represents the fulcrum for a new signalingpathway in which independent ligands, the Agouti protein and melanocortin pep-tides, inhibit each other’s binding and transduce opposite signals through a single re-ceptor.

3–10

In what follows, we review briefly the molecular genetics of Agoutiprotein and Agouti-related protein (Agrp), highlight recent studies from our labora-tory that bear on their biochemical mechanism of action, and present data that impli-cate additional gene products in Agouti and melanocortin receptor signaling.

a

Address for Correspondence: Greg Barsh, Beckman Center B271A, Stanford UniversitySchool of Medicine, Stanford, CA 94305-5323, USA. 650-723-5035 (voice); 650-723-1399(fax); gbarsh@cmgm.stanford.edu (e-mail).

144 ANNALS NEW YORK ACADEMY OF SCIENCES

MOLECULAR GENETICS OF AGOUTI AND AGOUTI-RELATED PROTEIN

The

Agouti

gene takes its name from the native language of a South AmericanIndian tribe where these rodents, also know as the Paca, exhibit a stereotypic coatcolor pattern in which individual hairs display a subapical band of yellow pigmenton an otherwise black background. This pattern develops dynamically during theanagen phase of hair growth, when melanocytes at the base of each follicle brieflyswitch from synthesizing black pigment, so-called eumelanin, to yellow pigment(pheomelanin), then back to eumelanin again as anagen hair growth is completed.

11

Most

Agouti

alleles affect only hair color, but several have dominant pleiotropic ef-fects that include a completely yellow coat, obesity, increased linear growth, prema-ture infertility, and immune abnormalities.

12,13

The prototype for obesity-associated

Agouti

alleles,

lethal yellow

(

A

y

), is one of the oldest known mouse mutations. It dif-fers from other obesity-associated

Agouti

alleles in that homozygosity for

A

y

causesembryonic death around the time of implantation.

5,14

Agouti

alleles associated with yellow pigment production are generally dominantto those associated with black pigment production. The dominant-recessive relation-ships apply to individual hair follicles and not the entire animal. For example,

black-and-tan

(

a

t

/

a

t

) mice, which have yellow ventral hairs and black dorsal hairs, crossedwith

Agouti

(

A

/

A

) mice, which have banded hairs on both ventrum and dorsum, pro-duce

A

/

a

t

mice, which have yellow ventral hairs and banded dorsal hairs—the so-called light-bellied Agouti phenotype. The genetic relationships between different

Agouti

alleles suggest that production of black pigment was a “default” state inwhich Agouti protein activate yellow pigment synthesis.

1

In addition, elegant trans-planation studies carried out by Silvers and colleagues more than thirty yearsago

15,16

indicate that

Agouti

caused cells in the dermis to release a paracrine mole-cule that act on nearby melanocytes in the epidermal portion of hair follicles.

These predictions from early genetic and embryologic studies were confirmedwhen

Agouti

was cloned and found to encode a 131-amino-acid secreted protein inwhich two domains were predicted: a basic region of approximately 60 amino acidsand a cysteine-rich carboxy-terminal region of approximately 40 amino acids

17,18

(see F

IGURE

1). Expression of

Agouti

RNA is normally limited to dermal papillaecells and its timing correlates with the synthesis of yellow pigment; i.e. follicles withbanded hairs express

Agouti

RNA at postnatal days 4–7, midway through anagen,whereas follicles with hairs that are completely yellow express

Agouti

RNA through-out the entire hair growth cycle.

19,20

Transgenic studies confirmed that

Agouti

had avery limited sphere of action, possibly because the highly charged amino-terminuslimited diffusion through the extracellular matrix (F

IG

. 1). Surprisingly,

A

y

and otherobesity-associated

Agouti

alleles were found to be caused by genomic rearrange-ments leading to ubiquitous expression of one or more transcripts predicted to en-code the normal open reading frame.

5,6,21

The ability of Agouti protein to elicitnonpigmentary phenotypes in alleles such as

A

y

is likely explained by the recent dis-covery of Agrp, whose RNA is expressed mainly in the hypothalamus and the adre-nal gland.

3,4

Although Agrp exhibits less than 20% identity with Agouti proteinoverall, it has the exact same size and genomic structure, and a nearly identical spac-ing of cysteine residues in the carboxy terminus (F

IG

. 1).

145BARSH et

al.

: MOLECULAR PHARMACOLOGY OF AGOUTI

FIGURE 1. Molecular genetics and pharmacology of Agouti and Agrp. A. Agoutiprotein and Agrp are of identical size and exon structure, but primary sequence similaritybetween the two protein is restricted to the cysteine-rich carboxy-terminal region, wherethere are 10 cysteines with a characteristic pattern of spacing that is also found in plectox-ins. Agouti, Agrp, and plectoxins have a cleavable signal sequence; however, plectoxinsundergo additional processing at the amino- and carboxy-termini to yield the cysteine-richsegment as the active molecule in vivo. B. In most animals, expression of Agouti protein isrestricted to the hair follicle where it causes melanocytes to produce yellow pigment via aparacrine mechanism. In mutant animals carrying Ay or similar mutations, ubiquitous ex-pression of transcripts that encode a normal Agouti protein cause pleiotropic effects due tothe ability of Agouti protein to mimic Agrp. C. Comparison of Agouti protein to Agrp intheir ability to inhibit α-MSH-induced accumulation of cAMP shows that Agouti proteinis a potent antagonist of the human Mc1r, Mc2r, and Mc4r, but has little effect on the Mc3ror Mc5r. By contrast, Agrp is a potent antagonist only at the Mc3r and Mc4r. Experimentssupporting these findings are described in References 3, 17, and 21.

146 ANNALS NEW YORK ACADEMY OF SCIENCES

BIOCHEMICAL MECHANISM OF AGOUTI SIGNALING

The similarity in cysteine spacing between Agouti protein and Agrp is also sharedby plectoxins and, to a lesser extent, omega-conotoxins, that are used by primitivehunting spiders and cone snails, respectively, to immobilize their prey.

22,23

Becausethese molecules are thought to act by inhibiting neuronal voltage-gated calciumchannels, initial speculation regarding the mechanism of Agouti protein action fo-cused on the possibility of a specific Agouti receptor that modulated calcium flux.

9,24

Although there is limited experimental support for a direct action of Agouti proteinon calcium channels,

8,25,26

most evidence indicates that Agouti protein and Agrp re-quire melanocortin receptors for signal transduction.

27–32

This family of G-protein-coupled receptors was first recognized by the ability of small circulating peptides,such as alpha-melanocyte stimulating hormone (

α

-MSH) or adrenocorticotrophichormone (ACTH), to activate adenylate cyclase in pigment cells or the adrenal cor-tex, respectively. More recently,

α

-MSH was recognized as an important neurotrans-mitter in the regulation of stereotypic behavior, central pathways governing theinflammatory response, and feeding.

33–35

A turning point in understanding the biochemical mechanism of Agouti proteinaction stems from the observation of Geschwind that administration of

α

-MSH to

A

y

animals causes a switch of pigment synthesis from yellow to black;

36,37

in hindsight,Agouti protein action is clearly opposite that of

α

-MSH. This opposite relationshipwas underscored by the recent observation of Cone and colleagues,

38

that the

reces-sive yellow

mutation, in which animals display mostly yellow pigment, is caused bya loss-of-function in the melanocyte receptor for

α

-MSH, the melanocortin-1 recep-tor (Mc1r).

These observations set the stage for pharmacologic studies demonstrating thatAgouti protein (and later Agrp) would inhibit activation of adenylate cyclase by mel-anocortin peptides.

27–30,39,40

However, the exact mechanism of Agouti and Agrp ac-tion has been difficult to unravel, in part because direct binding assays for Agoutihave been difficult to develop. We recently developed a direct binding assay for re-combinant Agouti protein in which an epitope-tagged form of the molecule (HA-Agouti) was first incubated with heterologous cells engineered to express the Mc1r,then, after a brief wash and fixation, cell surface binding of the epitope was detectedimmunohistochemically.

32

Using this so-called “overlay” assay, we found that

α

-MSH or other melanocortin agonists would inhibit cell surface binding of HA-Agouti to cells that express the Mc1r (see F

IGURE

2A). These observations, and theirconverse,

27,28,30

argue strongly that Agouti protein is a competitive antagonist ofmelanocortin peptides, but do not distinguish between competition for the identicalsite versus induction of a conformational change that prevents binding of the alter-native ligand, so-called allosteric competition. The carboxy-terminal portion of Ag-outi or Agrp is sufficient for melanocortin receptor binding,

3,32,39

and it is likely tofold into a compact globular structure with five pairs of intrachain disulfide bonds ina 40-amino-acid segment. By contrast, the active core of melanocortin agonists is aseven-residue, relatively hydrophobic, peptide that is thought to interact with themembrane-spanning portion of the receptor. The question of whether Agouti proteinand

α

-MSH bind to the same, different, or overlapping sites is of practical as well as

147BARSH et

al.

: MOLECULAR PHARMACOLOGY OF AGOUTI

theoretical importance, because small molecule pharmaceutical agents targeted to abinding site specific for Agouti protein or Agrp might not have the same physiologiceffects as melanocortin peptide mimetics.

Pharmaologic studies can sometimes provide insight into mechanisms of receptorantagonism, since increasing concentrations of an antagonist that competes for thesame site as agonist will cause a proportionate and parallel rightward shift of the ag-onist dose-response curve. Most of our work on Agouti protein pharmacology hasutilized the endogenous melanocortin receptor present on a

Xenopus

pigment cell

FIGURE 2. Mechanism of Agouti protein signaling in vitro and in vivo. A. Epitope-tagged full-length Agouti protein binds directly to the Mc1r and can be displaced either byexcess melanocortin peptide or by the cysteine-rich carboxy-terminal fragment of Agoutiprotein. However, pharmacologic studies of full-length Agouti protein on Xenopus melan-ophores indicate a time- and temperature-dependent potentiation of melanocortin inhibitionthat is likely explained by receptor downregulation or desensitization. B. In vivo, Agouti in-hibits expression of tyrosinase, and low levels of tyrosinase are associated with synthesis ofyellow pigment. In a genetic background, where tyrosinase activity is compromised becauseof the chinchilla (Tyrch) mutation, expression of Agouti causes a further reduction in tyro-sinase activity such that very little pigment of any type can be made, leading to a cream col-or. The cream-colored phenotype observed in Ay; chinchilla mice is clearly distinct from thepale yellow phenotype observed in Mc1re/Mc1re; chinchilla mice, and suggests that Agouticauses a greater inhibition of tyrosinase activity than does loss-of-function at the Mc1r. Ex-periments supporting these findings are described in References 32 and 42.

148 ANNALS NEW YORK ACADEMY OF SCIENCES

line, melanophores, developed by Dr. Michael Lerner.

41

This assay system is ex-tremely sensitive, robust, and offers an opportunity to study the effects of melano-cortins or Agouti protein in real time. We find that the effects of Agouti proteincannot be explained solely by agonist displacement, since full-length Agouti proteinexhibts a time- and temperature-dependent potentiation of melanocortin inhibitionthat suggests a direct effect of Agouti protein on receptor downregulation

32

(F

IG

. 2A). For example, increasing concentrations of Agouti protein cause a reduc-tion in the maximum level of

α

-MSH–induced pigment dispersion, and the level ofreduction depends on time, temperature, and Agouti protein concentration. Surpris-ingly, these noncompetitive effects are observed only with intact full-length Agoutiprotein.

42

The carboxy-terminal cysteine-rich domain by itself, or full-length Agoutiprotein treated with a protease that cleaves between the amino- and carboxy-terminaldomains, have no effect on the maximum level of

α

-MSH–induced pigment disper-sion.

We have also used pigmentation genetics to examine the relationship betweenAgouti protein signaling and Mc1r function

in vivo

.

32

Decreased Mc1r signaling inmelanocytes, whether caused by increased Agouti expression or decreased Mc1r ex-pression, leads to reduced activity of tyrosinase, a rate-limiting enzyme for synthesisof all pigment types.

43,44

On an appropriate genetic background provided by thechinchilla (Tyrch) mutation, Ay/a mutant mice exhibit a greater reduction of tyrosi-nase activity then Mc1re/Mc1re mutant mice, leading to an obvious coat color differ-ence, cream-colored versus yellow32 (FIG. 2B). Animals, mutant for both mutations(Ay/a and Mc1re/Mc1re), exhibit a pigmentation phenotype identical to the singleMc1re/Mc1re mutants; therefore, the ability of Agouti expression to depress tyrosi-nase activity below a level associated with Mc1r deficiency requires a functionalMc1r receptor. Taken together with the in vitro studies, these findings suggest thatAgouti protein transduces a signal via the Mc1r independent of that caused byα-MSH displacement.

OTHER COAT COLOR GENES AND THE AGOUTI SIGNALING PATHWAY

Many of the critical tools and key insights for understanding the pathways de-scribed above are based on previously existing mutations, recognized by their effectson pigmentation. Two additional coat color mutations that provide insight into Ago-uti signaling are mahogany and mahoganoid.45 Each produces a similar phenotype;an increased amount of black/brown pigment, eumelanin, compared to red/yellowpigment, pheomelanin (see FIGURE 3). These effects are similar to those caused byloss-of-function Agouti mutations or constitutively activating Mc1r mutations and, apriori, might be explained by decreased production or cell surface binding of Agoutiprotein, increased production of α-MSH, or defects in the melanocyte enzymes orstructural proteins required to produce yellow pigment granules. Plasma levels ofα-MSH and ACTH are not altered by the mahogany or mahoganoid mutations;46 weand others47,48 have used double mutant studies to distinguish among the other al-ternatives. If mahogany were caused by defects in the melanocyte machinery re-quired for yellow pigment synthesis, one would expect mahogany to suppress the

149BARSH et al.: MOLECULAR PHARMACOLOGY OF AGOUTI

pigmentation phenotype of recessive yellow, a loss of function mutation of the Mc1r.However, animals doubly mutant for mahogany and recessive yellow are indistin-guishable from single mutant recessive yellow animals.46 Similarly, if mahoganywere caused by reduced transcription of Agouti RNA in the hair follicle, one wouldexpect that mahogany should have no effect on the yellow coat color caused by Ay,resulting from a deletion that places Agouti RNA under control of a ubiquitously ex-pressed promoter. However, animals doubly mutant for Ay and mahogany exhibit adarkened coat color.46 Thus, mahogany and mahoganoid lie downstream of Agoutitranscription but upstream of the Mc1r, and therefore may be required for posttrans-lational processing, secretion, or binding of Agouti protein to the cell surface(FIG. 3).

ACKNOWLEDGMENTS

We are grateful to Chris Kaelin and Julie Kerns for their help and criticism. Thiswork was supported in part by Grant DK28506 from the National Institutes ofHealth. G.S.B. is an Associate Investigator of the Howard Hughes Medical Institute.

FIGURE 3. Genetic pathway for Agouti signaling. Mahogany (mg) and mahoganoid(md) are each required for Agouti signaling, since presumptive loss-of-function mutationsin either gene cause a phenotype similar to that caused by a loss-of-function Agouti muta-tion. Double mutant studies, as described in the text and in the references, place mg and mddownstream of Agouti transcription but upstream of the Mc1r. Experiments supportingthese findings are described in Reference 46.

150 ANNALS NEW YORK ACADEMY OF SCIENCES

REFERENCES

1. SILVERS, W.K. 1979. The Agouti and Extension Series of Alleles, Umbrous and Sable.6–44. Springer-Verlag, New York.

2. JACKSON, I.J. 1994. Molecular and developmental genetics of mouse coat color. Annu.Rev. Genet. 28: 189–217.

3. OLLMANN, M.M., B.D. WILSON, Y.K. YANG, et al. 1997. Antagonism of central mel-anocortin receptors in vitro and in vivo by agouti-related protein. Science 278: 135–138.

4. SHUTTER, J.R., M. GRAHAM, A.C. KINSEY, S. SCULLY, R. LUTHY & K.L. STARK. 1997.Hypothalamic expression of ART, a novel gene related to agouti, is up-regulated inobese and diabetic mutant mice. Genes Dev. 11: 593–602.

5. DUHL, D.M.J., M.E. STEVENS, H. VRIELING et al. 1994. Pleiotropic effects of themouse lethal yellow (A(y)) mutation explained by deletion of a maternally expressedgene and the simultaneous production of agouti fusion RNAs. Development 120:1695–1708.

6. MICHAUD, E.S., S.J. BULTMAN, M.L. KLEBIG et al. 1994. A molecular model for thegenetic and phenotypic characteristics of the mouse lethal yellow (Ay) mutation.Proc. Natl. Acad. Sci. USA 91: 2562–2566.

7. MILTENBERGER, R.J., R.L. MYNATT, J.E. WILKINSON & R.P. WOYCHIK. 1997. The roleof the agouti gene in the yellow obese syndrome. J. Nutr. 127: 1902S–1907S.

8. Michaud, J., R.L. MYNATT, R.J. MILTENBERGER et al. 1997. Role of the agouti gene inobesity. J. Endocrinol. 155: 207-209.

9. MANNE, J., A.C. ARGESON & L.D. SIRACUSA. 1995. Mechanisms for the pleiotropiceffects of the agouti gene. Proc. Natl. Acad. Sci. USA 92: 4721–4724.

10. PERRY, W.L., N.G. COPELAND & N.A. JENKINS. 1994. The molecular basis for domi-nant yellow agouti coat color mutations. Bioessays 16: 705–707.

11. SEARLE, A.G. 1968. Comparative Genetics of Coat Color in Mammals. AcademicPress, New York.

12. YEN, T.T., A.M. GILL, L.G. FRIGERI, G.S. BARSH & G.L. WOLFF. 1994. Obesity, dia-betes, and neoplasia in yellow A(vy)/- mice: ectopic expression of the agouti gene.Faseb. J. 8: 479–488.

13. WOLFF, G.L., D.W. ROBERTS & D.B. GALBRAITH. 1986. Prenatal determination of obe-sity, tumor susceptibility, and coat color pattern in viable yellow (Avy/a) mice. Theyellow mouse syndrome. J. Heredity 77: 151–158.

14. CUÉNOT, L. 1905. Les races pures et leurs combinasisons chez les souris (4me note).Arch. Zool. Exp. Gen. 4e Ser 3: cxxiii–cxxxii.

15. POOLE, T.W. 1980. Dermal-epidermal interactions and the action of alleles at the ago-uti locus in the mouse. II. The viable yellow (Avy) and mottled agouti (am) alleles.Dev. Biol. 80: 495–500.

16. POOLE, T.W. 1975. Dermal-epidermal interactions and the action of alleles at the ago-uti locus in the mouse. Dev. Biol. 42: 203–210.

17. MILLER, M.W., D.M.J. DUHL, H. VRIELING et al. 1993. Cloning of the mouse agoutigene predicts a secreted protein ubiquitously expressed in mice carrying the Lethal-Yellow mutation. Genes Dev. 7: 454–467.

18. BULTMAN, S.J., E.J. MICHAUD & R.P. WOYCHIK. 1992. Molecular characterization ofthe mouse agouti locus. Cell 71: 1195–1204.

19. Millar, S.E., M.W. MILLER, M.E. STEVENS & G.S. BARSH. 1995. Expression and trans-genic studies of the mouse agouti gene provide insight into the mechanisms bywhich mammalian coat color patterns are generated. Development 121: 3223–3232.

20. VRIELING, H., D.M. DUHL, S.E. MILLAR, K.A. MILLER & G.S. BARSH. 1994. Differ-ences in dorsal and ventral pigmentation result from regional expression of themouse agouti gene. Proc. Natl. Acad. Sci. USA 91: 5667–5671.

21. DUHL, D.M.J., H. VRIELING, K.A. MILLER, G.L. WOLFF & G.S. BARSH. 1994. Neomor-phic agouti mutations in obese yellow mice. Nat. Genet. 8: 59–65.

151BARSH et al.: MOLECULAR PHARMACOLOGY OF AGOUTI

22. QUISTAD, G.B. & W.S. SKINNER. 1994. Isolation and sequencing of insecticidal pep-tides from the primitive hunting spider, Plectreurys tristis (Simon). J. Biol. Chem.269: 11098–11101.

23. OLIVERA, B.M., J. RIVIER, J.K. SCOTT, D.R. HILLYARD & L.J. CRUZ. 1991. Conotox-ins. J. Biol. Chem. 266: 22067–22070.

24. ZEMEL, M.B., J.H. KIM, R.P. WOYCHIK et al. 1995. Agouti regulation of intracellularcalcium: role in the insulin resistance of viable yellow mice. Proc. Natl. Acad. Sci.USA 92: 4733–4737.

25. JONES, B.H., J.H. KIM, M.B. ZEMEL et al. 1996. Upregulation of adipocyte metabolismby agouti protein: possible paracrine actions in yellow mouse obesity. Amer. J.Physiol.-Endocrinol. Met. 33: E192–E196.

26. XUE, B., N. MOUSTAID-MOUSSA, W.O. WILKISON & M.B. ZEMEL. 1998. The agoutigene product inhibits lipolysis in human adipocytes via a Ca2+-dependent mecha-nism. Faseb. J. 12: 1391–1396.

27. LU, D.S., D. WILLARD, I.R. PATEL et al. 1994. Agouti protein is an antagonist of themelanocyte-stimulating-hormone receptor. Nature 371: 799–802.

28. BLANCHARD, S.G., C.O. HARRIS, O.R.R. ITTOOP et al. 1995. Agouti antagonism ofmelanocortin binding and action in the B16F10 murine melanoma cell line. Bio-chem. 34: 10406–10411.

29. SIEGRIST, W., R. DROZDZ, R. COTTI, D.H. WILLARD, W.O. WILKISON & A.N. EBERLE.1997. Interactions of alpha-melanotropin and agouti on B16 melanoma cells: evi-dence for inverse agonism of agouti. J. Receptor Signal Transd. Res. 17: 75–98.

30. YANG, Y.K., M.M. OLLMANN, B.D. WILSON et al. 1997. Effects of recombinant ago-uti-signaling protein on melanocortin action. Mol. Endocrinol. 11: 274–280.

31. SIEGRIST, W., D.H. WILLARD, W.O. WILKISON & A.N. EBERLE. 1996. Agouti proteininhibits growth of b16 melanoma cells in vitro by acting through melanocortinreceptors. Biochem. Biophys. Res. Commun. 218: 171–175.

32. OLLMANN, M.M., M.L. LAMOREUX, B.D. WILSON & G.S. BARSH. 1998. Interaction ofAgouti protein with the melanocortin 1 receptor in vitro and in vivo. Genes Dev. 12:316–330.

33. O’DONOHUE, T.L. & D.M. DORSA. 1982. The opiomelanotropinergic neuronal andendocrine systems. Peptides 3: 353–395.

34. DE WIED, D. 1993. Melanotropins as neuropeptides. Ann. N.Y. Acad. Sci. 680: 20–28.35. CONE, R.D., D. LU, S. KOPPULA et al. 1996. The melanocortin receptors: agonists,

antagonists, and the hormonal control of pigmentation. Recent Prog. Horm. Res. 51:287–318.

36. GESCHWIND, I.I. 1966. Change in hair color in mice induced by injection of α-MSH.Endocrinol. 79: 1165–1167.

37. GESCHWIND, I.I., R.A. HUSEBY & R. NISHIOKA. 1972. The effect of melanocyte-stimulating hormone on coat color in the mouse. Rec. Prog. Horm. Res. 28: 91–130.

38. ROBBINS, L.S., J.H. NADEAU, K.R. JOHNSON et al. 1993. Pigmentation phenotypes ofvariant extension locus alleles result from point mutations that alter MSH receptorfunction. Cell 72: 827–834.

39. WILLARD, D.H., W. BODNAR, C. HARRIS et al. 1995. Agouti structure and function:characterization of a potent alpha-melanocyte stimulating hormone receptor antago-nist. Biochem. 34: 12341–12346.

40. KIEFER, L.L., J.M. VEAL, K.G. MOUNTJOY & W.O. WILKISON. 1998. Melanocortinreceptor binding determinants in the agouti protein. Biochem. 37: 991–997.

41. POTENZA, M.N. & M.R. LERNER. 1992. A rapid quantitative bioassay for evaluatingthe effects of ligands upon receptors that modulate cAMP levels in a melanophorecell line. Pigment Cell. Res. 5: 372–378.

152 ANNALS NEW YORK ACADEMY OF SCIENCES

42. OLLMANN, M. 1998. Mechanisms of Agouti protein signaling. Ph.D. Thesis, StanfordUniversity School of Medicine, Stanford.

43. KOBAYASHI, T., W.D. VIEIRA, B. POTTERF, C. SAKAI, G. IMOKAWA & V.J. HEARING.1995. Modulation of melanogenic protein expression during the switch from eu- topheomelanogenesis. J. Cell Sci. 108: 2301–2309.

44. PROTA, G., M.L. LAMOREUX, J. MULLER et al. 1995. Comparative analysis of melaninsand melanosomes produced by various coat color mutants. Pigm. Cell Res. 8: 153–163.

45. LANE, P.W. & M.C. GREEN. 1960. Mahogany, a recessive color mutatioin in linkagegroup V of the mouse. J. Hered. 51: 228–230.

46. MILLER, K.A., T.M. GUNN, M.M. CARRASQUILLO, M.L. LAMOREUX, D.B. GALBRAITH

& G.S. BARSH. 1997. Genetic studies of the mouse mutations mahogany and mahog-anoid. Genetics 146: 1407–1415.

47. BEECHEY, C.V. & A.G. SEARLE. 1978. Epistasy of extension alleles. Mouse News Lett.59: 19.

48. BEECHEY C.V. & A.G. SEARLE. 1979. Epistasy of nc over Ay. Mouse News Lett. 60:45–47.